Speech Synthesis Technology: Status and Challenges

Electrical and Computer Engineering, College of Engineering, University of Massachusetts Amherst, Amherst, Massachusetts 01003, United States of America

^* Corresponding author: caiyuechen@umass.edu

Abstract

In recent years, speech synthesis technology has been more widely used in the field of artificial intelligence and human-computer interaction due to the excess of machine learning models to deep learning models. With the rise and development of applications such as intelligent voice assistants, voice navigation systems, generative macro modelling, and virtual reality, Users’ demand for voice systems is not limited to the generated sound as cold as robots and full of “inhuman” tones and rhythm, it is also desired to generate speech that is more natural, fluent, and free of mechanical sensations. This paper reviews the recent development of speech synthesis techniques and the interpretation of each step in the order of speech synthesis steps, two parts of acoustic modeling and wave pattern synthesis are described in detail. In addition, this article aims to introduce some typical Speech Synthesis technology and also summarizes the current applications and future prospects in the field of speech synthesis research.